PhD Student • Computer Science • NMSU Las Cruces, NM

Building reliable ML for security & data-scarce settings.

I’m Jiefei Liu. My work spans Generative AI (Diffusion / GAN / LLMs), Federated Learning, Continual/Class-Incremental Learning, and Unseen Class Detection, with a focus on robust Intrusion Detection Systems (IDS).

Download CV Email Google Scholar GitHub LinkedIn
Email jiefei9657@gmail.com Phone 575-571-0810

About

Short bio + education

I build ML methods that stay reliable under distribution shift (new/unseen classes, limited labels, privacy constraints). I’ve worked on federated IDS, generative network traffic synthesis, and applied ML systems.

  • PhD, Computer Science — New Mexico State University (Aug 2022 – Present), anticipated graduation: 05/2027
  • MS, Computer Science — NMSU (Aug 2020 – Aug 2022)
  • BS, Computer Science — NMSU (Jan 2016 – May 2020), minors: EE & Mathematics

Focus areas

Generative AI (Diffusion, GANs, LLMs) Federated Learning Continual / Class-Incremental Unseen Class Detection Intrusion Detection Systems

Core stack

PyTorch NumPy / Pandas scikit-learn flwr Flask Linux / Git

Research

What I’m working on

Federated ML for Network Vulnerability Assessment and Monitoring (DoD)

Unseen class detection • OOD generation • federated class-incremental learning
  • Develop methods to address unseen class detection challenges in IDS, improving reliability against novel cyber threats.
  • Fine-tune LLMs to generate synthetic IDS datasets; compare with GAN and diffusion for quality, diversity, and realism under privacy/data scarcity constraints.
  • Explore prompt-based LLMs for direct attack classification and future unseen attack detection.
  • Design a robust federated class-incremental learning framework for scalable, adaptive FL.

Probing Attacks on Networks and Mitigation Using ML (DoD)

Federated IDS • imbalance • communication reduction
  • Design federated learning frameworks for IDS, enabling decentralized yet secure model training.
  • Address local class imbalance via data augmentation and realistic scenario simulation.
  • Achieve ≥2× performance improvement over baseline methods.
  • Evaluate GAN/diffusion within FL to reduce communication overhead while preserving privacy.
  • Achieve 96% reduction in communication cost compared to previously proposed FL framework.

Projects

Selected applied work

Cow Trajectory Analysis Project

Time-series segmentation • hierarchical clustering • behavior labeling
  • Applied time-series segmentation (ClaSP) and hierarchical clustering to group GPS-tracked data points with similar patterns.
  • Enabled biologists to more efficiently label cow behaviors, supporting downstream animal behavior analysis.

Midcontinent Independent System Operator (MISO) Company Project

Log parsing • data verification • reliability analytics
  • Extracted data points from power grid alarm logs; performed verification, cleaning, and preprocessing for high-quality inputs.
  • Analyzed alarm data for reliability insights using Pandas, scikit-learn, and Matplotlib.

Constrained Skyline Queries (CSQ) over Transportation Networks

Web demo • interactive maps • server response visualization
  • Developed a web-based CSQ demo to submit queries, process server responses, and visualize paths/results on interactive maps.
  • Tech: Google Maps API, HTML, JavaScript.

Python/NLP-based Academic Voice Search System

Flask backend • NLP • topic modeling
  • Processed queries on a Flask backend; used NLTK for NLP and Gensim for topic modeling.
  • Returned and displayed the most relevant results on the frontend.

Publications

Full list

Under review & in preparation

Scholar

Diffusion-based Multi-Model Federated Learning for Network Intrusion Detection

Jiefei Liu, Huiping Cao, Abu Saleh Md Tayeen, Qixu Gong, Satyajayant Misra, Pratyay Kumar, Jayashree Harikumar. Submitted to IEEE Transactions on Networking.

NetDiffuser: Deceiving DNN-Based Network Attack Detection Systems with Diffusion-Generated Adversarial Traffic

Pratyay Kumar, Abu Saleh Md Tayeen, Satyajayant Misra, Huiping Cao, Jiefei Liu, Qixu Gong, Jayashree Harikumar. Submitted to IEEE Transactions on Information Forensics and Security.

Peer-reviewed conference & journal articles

10 items

Is Synthetic Flow Data from Generative Models Ready for Network Intrusion Detection Systems?

Jiefei Liu, Qixu Gong, Wenbin Jiang, Pratyay Kumar, Abu Saleh Md Tayeen, Huiping Cao, Satyajayant Misra, Jayashree Harikumar. IEEE MILCOM 2025.

NeTIF: Network Traffic to Image Features for Robust Intrusion Detection

Stephen Villanueva, Abu Saleh Md Tayeen, Qixu Gong, Satyajayant Misra, Aden Dogar, Huiping Cao, Jiefei Liu, Pratyay Kumar, Jayashree Harikumar. IEEE MILCOM 2025.

NetPrompt: Evaluation of LLMs as Network Intrusion Detection System

Pratyay Kumar, Abu Saleh Md Tayeen, Qixu Gong, Jiefei Liu, Satyajayant Misra, Huiping Cao, Jayashree Harikumar. IEEE MILCOM 2025.

Feature Selection via Class-wise Mean Deviation

Abu Fuad Ahmad, Jiefei Liu, Qixu Gong, Satyajayant Misra, Jayashree Harikumar. Accepted at ICMLA 2025.

Development of a Novel Classification Approach for Cow Behavior Analysis using Tracking Data and Unsupervised Machine Learning Techniques

Jiefei Liu, Derek W. Bailey, Huiping Cao, Tran Cao Son, Colin T. Tobin. MDPI Sensors, 2024.

Evaluation of Skyline Path Queries over Road Networks with Graph Neural Network Support

Qixu Gong, Huiying Chen, Huiping Cao, Jiefei Liu. ACM Transactions on Spatial Algorithms and Systems (TSLAS), 2024.

Multi-Model-based Federated Learning to Overcome Local Class Imbalance Issues

Jiefei Liu, Huiping Cao, Abu Saleh Md Tayeen, Satyajayant Misra, Pratyay Kumar, Jayashree Harikumar. ICMLA 2023, pp. 265–270.

FLNET2023: Realistic Network IDS Dataset for Federated Learning

Pratyay Kumar, Jiefei Liu, Abu Saleh Md Tayeen, Satyajayant Misra, Huiping Cao. IEEE MILCOM 2023, pp. 345–350.

Class-Specific Attention for Time-Series Classification

Yifan Hao, Huiping Cao, K. Selçuk Candan, Jiefei Liu, Huiying Chen, Ziwei Ma. arXiv:2211.10609, 2022.

CSQ System: Constrained Skyline Queries on Transportation Networks

Qixu Gong, Jiefei Liu, Huiping Cao. IEEE ICDE 2020, pp. 1746–1749.

Abstracts

2 items

Animal Behavior Analysis Using Unsupervised ML

Jiefei Liu, Derek W. Bailey, Huiping Cao, Tran Cao Son, Colin T. Tobin. Journal of Animal Science, 2023.

Prediction of Short-Term Drought Impacts Using ML: A Case Study for New Mexico

Hatim M.E. Geli, Lindsay E. Johnson, Michael J. Hayes, Huiping Cao, Jiefei Liu, Hasan Al-Qudah. AMS Annual Meeting, 2022.

Experience

Roles + skills

Research Assistant — New Mexico State University

May 2021 – Present • Las Cruces, NM

Federated IDS (DoD projects), generative traffic synthesis and evaluation, and applied ML projects.

Research Assistant — NMSU

Aug 2018 – Aug 2020 • Las Cruces, NM

Projects of predicte drought indices using Python ML.
Developed a web demo for constrained skyline queries (CSQ) over transportation networks.

Skills

Programming: Python, Java, R, HTML, JavaScript, C/C++ ML/AI: Federated & Continual Learning, Diffusion/GAN/LLM, Open-set/OOD Libraries: PyTorch, Pandas, NumPy, scikit-learn, Matplotlib, flwr, NLTK, Gensim Other: Linux, Git, Cloud/HPC, APIs (Google Maps/Voice)

Contact

Links

Direct

Email: jiefei9657@gmail.com
Phone: 575-571-0810
Google Scholar: scholar.google.com
GitHub: github.com/JiefeiLiu
LinkedIn: linkedin.com/in/jiefei-liu-877501266

Tip: keep Jiefei_s_CV.pdf next to index.html for the CV button to work.

Quick actions

Email me Open Scholar Open CV
© Jiefei Liu • Single-file site.